Method and system for bridge damage detection

Abstract

Methods and systems for bridge damage detection using, for example, one or more strain range methods are provided. One exemplary embodiment provides a computer-implemented methods and systems for determining bridge damage from strain sensor data, for example, by collecting a batch of strain data from one or more sensor pairs. From the batch of strain data one or more sets of strain data may be extracted comprising a quasi-static response of the bridge under ambient traffic loads. A relationship may be established between the one or more sets of strain data extracted from the one or more sensor pairs by orthogonal regression. Bridge damage may be detected by generally isolating a damage indicator between the one or more sensor pairs by monitoring changes in a statistical F shm value over time.

What is claimed is: 
     
       1. A computer-implemented method for determining bridge damage from strain sensor data, comprising:
 collecting a batch of strain data from one or more sensor pairs; 
 after collecting the batch of strain data, segmenting the batch of strain data to remove temperature variations in said batch of strain data to zero the batch of strain data, wherein said removal of temperature variations includes determining a temperate baseline for the data set and subtracting said temperate baseline from the segmented batch of strain data; 
 removing bridge component responses from the batch of strain data to acquire a quasi-static response; 
 extracting from the batch of strain data one or more sets of strain data comprising the quasi-static response of the bridge under ambient traffic loads; 
 establishing a relationship between the one or more sets of strain data extracted from the one or more sensor pairs by orthogonal regression; and 
 detecting bridge damage by generally isolating a damage indicator between the one or more sensor pairs using a statistical F shm  value, wherein the damage indicator comprises a change in the one or more F shm  values for a control chart. 
 
     
     
       2. The computer-implemented method of  claim 1  wherein zeroing the batch of strain data does not remove random noise or dynamic induced behaviors. 
     
     
       3. The computer-implemented method of  claim 1  further comprising:
 developing a control chart for each sensor in the one or more sensor pairs for tracking change in the statistical F shm  value. 
 
     
     
       4. The computer-implemented method of  claim 1  wherein isolating the damage indicator comprises limiting false-positives resulting from residuals in sensor pairs proximate bridge damage. 
     
     
       5. The computer-implemented method of  claim 1  further comprising:
 performing at least one training protocol for characterizing a baseline behavior for the one or more sensor pairs. 
 
     
     
       6. A computer-implemented system for determining bridge damage from strain sensor data, comprising:
 a database having a batch of strain data collected from one or more sensor pairs, said batch of strain data zeroed after collection by segmenting the batch of strain data to remove temperature variations, and wherein said removal of temperature variations includes determining a temperate baseline for the data set and subtracting said temperate baseline from the segmented batch of strain data; 
 one or more sets of strain data from the batch of strain data comprising a quasi-static response of the bridge under ambient traffic loads; 
 an orthogonal regression relationship between the one or more sets of strain data; and 
 a bridge damage indicator between the one or more sensor pairs detected by an isolation protocol using a statistical F shm  value; 
 wherein the quasi-static response of the bridge is independent of bridge component responses from the batch of strain by removing the bridge component responses from the batch of strain to acquire a quasi-static response; and 
 wherein the damage indicator comprises a change in the one or more F shm  values for a control chart. 
 
     
     
       7. The computer-implemented system of  claim 6  wherein zeroing the batch of strain data does not remove random noise or dynamic induced behaviors. 
     
     
       8. The computer-implemented system of  claim 6  further comprising:
 a control chart for each sensor in the one or more sensor pairs, wherein the control chart is based at least in part on the statistical F shm  value. 
 
     
     
       9. The computer-implemented system of  claim 6  wherein the bridge damage indicator comprises a limited false-positives from residuals in sensor pairs proximate bridge damage. 
     
     
       10. The computer-implemented system of  claim 6  further comprising:
 at least one training protocol for characterizing a baseline behavior for the one or more sensor pairs. 
 
     
     
       11. A computer-implemented method for determining bridge damage from strain sensor data under ambient traffic loads, comprising:
 collecting a batch of strain data from one or more sensor pairs attached to a bridge support structure; 
 segmenting the batch of strain data into time segments to remove the influence of any temperature variations to the batch of strain data, wherein said removal of the influence of temperature variations includes determining a temperate baseline for the data set and subtracting said temperate baseline from the segmented batch of strain data; 
 extracting from the batch of strain data one or more sets of strain data comprising a quasi-static response of the bridge independent of bridge component responses, wherein the bridge component responses is removed from the batch of strain to acquire the quasi-static response; 
 identifying vehicular events based on a statistical and structural evaluation of the quasi-static response relative to a location of the one or more sensor pairs; 
 establishing an orthogonal regression relationship between the one or more sets of strain data extracted from the one or more sensor pairs; 
 isolating a bridge damage indicator between the one or more sensor pairs using a statistical F shm  value from a control chart, wherein the damage indicator comprises a change in the one or more F shm  values for a control chart; and 
 detecting bridge damage from the bridge damage indicator independent of the bridge component responses. 
 
     
     
       12. The computer-implemented method of  claim 11  further comprising:
 developing the control chart for each sensor in the one or more sensor pairs based at least on part on the statistical F shm  value. 
 
     
     
       13. The computer-implemented method of  claim 11  further comprising:
 performing at least one training protocol for characterizing a baseline behavior for the one or more sensor pairs. 
 
     
     
       14. The computer-implemented method of  claim 11  further comprising:
 monitoring the control chart for tracking how the statistical F shm  value changes over time. 
 
     
     
       15. The computer-implemented method of  claim 11  wherein isolating the bridge damage indicator comprises limiting false-positives resulting from residuals in sensor pairs proximate bridge damage.